tudies conducted on Marine recruits inParris Island, South Carolina (15), :he patterning of such outbreaks, provide further evidence against the odox explanation. The basic training program lasts for 8 weeks. The .ber of upper respiratory infections increases from the first through the 1 weeks, decreases in the fourth through the sixth weeks, and begins to jase again in the seventh and eighth weeks. As far as can be determined, e are no differences in crowding during these 8 weeks. Furthermore, sick from all causes, including gastrointestinal, musculo-skeletal, skin infec-s, trauma, and so on, display a similar pattern. Not only is this regularity trved for all platoons, but there are systematic differences in the rate of :tion between platoons (living under identical conditions), some exhibit-i markedly higher rate for their entire 8 weeks than others, is will be indicated later in this paper, such data do not necessarily refute role of crowding in changing susceptibility to disease, but they do provide s which may necessitate a change in our thinking about how crowding can lence health.
it first sight the animal data appear convincing. However, they, too, need e re-examined both in terms of consistency and in terms of the extrapola-
that can be made to human populations. Kessler at the Rockefeller itute, for example, has indicated that mice under extreme conditions of vding exhibited no increase in pathology once the population had eved its maximum density and no further population growth was oc-ing. Under these circumstances asocial behavior was common but physical lology no more frequent than in the control group living under un-vded conditions. However, during the phase of rapid population growth
preceded this plateau, disease was much more frequent than in the trol group (16).
t is apparent even from these fragmentary illustrative data that population sity or crowding does not inevitably lead to poorer health. The rest of this er will examine some of the reasons that may account for the conflicting i and suggest the need to reformulate some of our conceptual models if effects of such phenomena are to be better understood, 'art of the reason for the discrepancies in the data presented earlier lies in well-recognized fact that many studies have used different and often lequate indicators of crowding. The indicators used have frequently been ble to distinguish between a high population density in some arbitrarily neated areas of land and increased social interaction. Second, crowding er certain circumstances may be associated with certain factors which nselves can influence health (poverty, poor nutrition, poor housing, etc.) under other circumstances may be associated with different factors. The